Inspiration
Attended a few Machine Learning workshops provided by Hack the North and we were inspired to create something similar ourselves!
What it does
Tracks hand movement on a webcam and maps it onto a canvas. Depending on the orientation of the hand, our website will predict the kind of hand gesture it most resembles.
How we built it
Used next.js and tensorflow.
Challenges we ran into
We came into this hackathon without any prior knowledge of machine learning and had big goals that we had to work around because of time. We had to downsize our goals a lot and make tradeoffs to make a minimum viable product.
Accomplishments that we're proud of
We got the tracking to work and were able to manually configure the Machine Learning Algorithm to guess some unique hand gestures not provided in the training dataset.
What we learned
Machine Learning and the impact that a training dataset can have on the accuracy of the algorithm and its ability to be applied in the real world.
What's next for Signal Guesser
We want it to be able to recognize the sign language alphabet with a decent accuracy.
Built With
- fingerpose
- javascript
- next.js
- tailwind
- tensorflow
Log in or sign up for Devpost to join the conversation.